HOM's
If you do a search on "phase plugs, compression drivers" you'll see a variety of gadgets meant to shape the wavefront coming out of a compression driver. It appears to me there is a lot of development going on to make the wavefront as close to planar as possible. It also appears to me that the circuitous route traveled by the wavefront to get to the mouth of the compression driver makes this a real challenge. It appears some of the energy exiting the compression driver will be headed right at the wall of the guide....into an early reflection. As Earl said, those reflections generate what he calls higher order modes.
It makes me wonder...what if the first part of the guide were absorptive, lossy? Would that reduce hom's? I'm beginning to wonder about a gradient of hard/lossy surfaces within the guide...and at the terminus...the gradient being tailored for best effect on the wave into the listening arena.
In your search you may also find links to gedlee's patents. I think Earl's being real nice to us here. 😉 I don't expect him to give away everything just yet.
If you do a search on "phase plugs, compression drivers" you'll see a variety of gadgets meant to shape the wavefront coming out of a compression driver. It appears to me there is a lot of development going on to make the wavefront as close to planar as possible. It also appears to me that the circuitous route traveled by the wavefront to get to the mouth of the compression driver makes this a real challenge. It appears some of the energy exiting the compression driver will be headed right at the wall of the guide....into an early reflection. As Earl said, those reflections generate what he calls higher order modes.
It makes me wonder...what if the first part of the guide were absorptive, lossy? Would that reduce hom's? I'm beginning to wonder about a gradient of hard/lossy surfaces within the guide...and at the terminus...the gradient being tailored for best effect on the wave into the listening arena.
In your search you may also find links to gedlee's patents. I think Earl's being real nice to us here. 😉 I don't expect him to give away everything just yet.
Creative plumbing I call it. Line arrays never get HF to sound right. But they pump it far away to be heard. Mission accomplished.
catapult said:
I agree with Dennis. With 6.5 degrees at the throat and a 60 degree exit angle, and a nominal throat radius of 1.0 the offset is X = 0.0379 at R = 1.0, and the radius at X = 0 is 0.9993. The math is pretty simple. There is no reason to do this iteratively. It's a direct calculation given the throat radius the throat angle and the exit (or flare) angle. Increaseing the throat angle to even 20 degrees with a 60 degree exit angle gives and offset of X = 0.121 and a radius at X = 0 of 0.97767.
john k... said:
Hey John,
I played a bit more and the difference does get bigger as you narrow the cone so that's probably what Earl is talking about. For a 6.5 degree throat, a 90 degree waveguide (theta=45) gives a diameter at X=0 of .9935 and theta=30 gives .9799. Still pretty tight tolerences for something carved out of wood or molded out of fiberglass but the difference is bigger than the .001 I mentioned before.
I wonder if a domed mid might be the closest to ideal for putting in a waveguide, since the wavefront it launches already has a shape close to perpendicular to the surface of the WG. Isn't that the purpose of the throat transition to the WG, to get the wave itself into the proper shape?
The only problem is that, no matter the shape of the diaphragm--concave, convex, or flat, it cannot create point-source radiation (a spherical wavefront) unless it is acoustically small or its output is constricted to an area that is acoustically small, as in a compression driver's throat.
Bill F. said:
While I realize it's not creating a point source, wouldn't the waveform be closer to the correct shape than the concave coned drivers I currently use, or is my cone flexing enough that the wave is decently shaped before it leaves the cone? If not, how far does it travel before the transition to a spherical shape occurs.
I believe the goal and function of the OS waveguide throat profile is to accept an (ideally) planar wavefront that is acoustically small WRT its wavelength and facilitate its transition into a hemisphical one.
So, actually, if you want to use a non-compression driver in an OS waveguide, a flat diaphragm would be closest to ideal.
I believe that loading the waveguide with a domed-diaphragm tweeter or a concave cone would be similar to using a compression driver that didn't launch a planar wave, and increased HOMs and a poorer pattern would be the result.
With a soft dome, which functions non-pistonically, or with a flexy cone, it would take some FEAing to figure out exactly what shape of wave it was launching at any given frequency.
In all cases, the driver's output should enter the waveguide through an acoustically small aperature.
So, actually, if you want to use a non-compression driver in an OS waveguide, a flat diaphragm would be closest to ideal.
I believe that loading the waveguide with a domed-diaphragm tweeter or a concave cone would be similar to using a compression driver that didn't launch a planar wave, and increased HOMs and a poorer pattern would be the result.
With a soft dome, which functions non-pistonically, or with a flexy cone, it would take some FEAing to figure out exactly what shape of wave it was launching at any given frequency.
In all cases, the driver's output should enter the waveguide through an acoustically small aperature.
Bill F. said:
I tried a quadratic throat with some 4" drivers, but I'm getting better sounding results with no throat at all. If the OS profile is all part of getting the correct wave shape with a compression driver, then maybe sticking with conical makes more sense with a coned driver, and a domed driver would seem even better.
catapult said:
Well the "exact" numbers are very slightly different that yours, 0.9934881 and 0.9803316, but who's counting. 🙂 So we still agree! I don't know why Earl seemed to indicate the math was "difficult".
Maybe because I was doing a 1" diameter and you were doing a 1" radius? As well, I just told Goal Seek to look for a 6.5 degree angle. It will give a more accurate result if you tell it to seek something like 6.5000 but that's getting really anal. 😉
john k... said:
John,
are you building something with a "waveguide" or is this just an engineering exercise?
cheers,
AJ
johninCR said:
An OS waveguide takes a planar wavefront and converts it to a spherical wavefront with the smoothest transition possible, therefore minimizing HOMs. However, the volume velocity across the plane is not uniform across the planar wavefront with commercially available drivers, so some HOMs are generated. This is why Earl went to the foam HOM absorber, because commercially availble drivers do not have a uniform volume velocity, because of the design of the phase plug. Earl designed for the lowest HOMs, and absorbs what is left. The foam is much cheaper than getting some company to redesign a phase plug. We can only hope that this aspect of phase plug design will be addressed in Thailand!!
A conical horn needs to be fed with a SPHERICAL wavefront to keep the HOMs away. Apparently, there are no drivers that will do this.
pooge said:
Even Earl's WG's don't get completely rid of HOM's. That's why he came up with the foam plugs. My cheap imitation uses foam as a lining on the waveguide surfaces, and it works well. The driver has the same sound as without a waveguide, but the uptilted response is gone and beaming HF beaming is much less noticeable. I thought an OS expansion might make further refinements and reduce size, but it sounds like it's not necessary without compression drivers.
My results make me a firm believer in Dr. Geddes approach, although I did add an OB twist.
AJinFLA said:
Well I don't really want to get involved in this thread. I just wanted to help out with the OS calculations. As of now I am not working with wave guides. They caught my interest as a means of offsetting a tweeter for a TP (ICTA) design, but I think there may be other problems with that application. However, you can not fill a thimble with what I know about wave guides applied to loudspeakers, at least today. 🙂
john k... said:
Me neither, but that didn't stop me from building one a while back and proclaiming it to be so 😀
Still working on the ICTA I see. Have you looked at the Duelund concept?
BTW, it would have been nice to see GD include with your Gradient filters. Good reading anyhow.
cheers,
AJ
AJinFLA said:
Not until you just mentioned it. Google it. I guess that is what they are doing, WG to align the tweeter. A pretty obvious solution to the offset problem, but their design is a good example of why I am not really crazy about the idea. In any (analog) TP design there is going to be considerable overlap in the crossover region so driver separation should be kept to a minimum, IMO. Actually, driver separation is one thing I don't like about the use of WGs in any case. Any to be honest I'm not exactly sure what problems a WG on a tweeter really solves other than offset.
As for GD, it's as expected. Goes up as the roll off order goes up. I'm adding more info on applying Eq to woofers with Qts greater than 0.5 and Fs that is acceptable, that is, where shifting Fs or the poles of a low Q response aren't necessary.
Can someone comment on the choice of a mid/woofer when using a waveguide for the benefit of other hobbyists such as myself who do not have the appropriate technical knowledge.
DDS has an excellent 90 degree waveguide (ENG1-90) that has a 10” mouth, a cutoff frequency of 900hz, a 1” throat and is 3” in depth.
If one has a mid/woofer in a 2 way that can extend to a high enough frequency without breakup, using a 2nd order crossover at 1,700 hz or a 3rd order at 1,300 hz, can you choose a driver that is either 8”, 10” or 12”?
Thanks,
Don
DDS has an excellent 90 degree waveguide (ENG1-90) that has a 10” mouth, a cutoff frequency of 900hz, a 1” throat and is 3” in depth.
If one has a mid/woofer in a 2 way that can extend to a high enough frequency without breakup, using a 2nd order crossover at 1,700 hz or a 3rd order at 1,300 hz, can you choose a driver that is either 8”, 10” or 12”?
Thanks,
Don